You are here
Long-term mortality after blood pressure-lowering and lipid-lowering treatment in patients with hypertension in the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) Legacy study: 16-year follow-up results of a randomised factorial trial
In patients with hypertension, the long-term cardiovascular and all-cause mortality effects of different blood pressure-lowering regimens and lipid-lowering treatment are not well documented, particularly in clinical trial settings. The Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) Legacy Study reports mortality outcomes after 16 years of follow-up of the UK participants in the original ASCOT trial.
ASCOT was a multicentre randomised trial with a 2 × 2 factorial design. UK-based patients with hypertension were followed up for all-cause and cardiovascular mortality for a median of 15·7 years (IQR 9·7–16·4 years). At baseline, all patients enrolled into the blood pressure-lowering arm (BPLA) of ASCOT were randomly assigned to receive either amlodipine-based or atenolol-based blood pressure-lowering treatment. Of these patients, those who had total cholesterol of 6·5 mmol/L or lower and no previous lipid-lowering treatment underwent further randomisation to receive either atorvastatin or placebo as part of the lipid-lowering arm (LLA) of ASCOT. The remaining patients formed the non-LLA group. A team of two physicians independently adjudicated all causes of death.
Of 8580 UK-based patients in ASCOT, 3282 (38·3%) died, including 1640 (38·4%) of 4275 assigned to atenolol-based treatment and 1642 (38·1%) of 4305 assigned to amlodipine-based treatment. 1768 of the 4605 patients in the LLA died, including 903 (39·5%) of 2288 assigned placebo and 865 (37·3%) of 2317 assigned atorvastatin. Of all deaths, 1210 (36·9%) were from cardiovascular-related causes. Among patients in the BPLA, there was no overall difference in cardiovascular mortality between treatments (adjusted hazard ratio [HR] 0·90, 95% CI 0·81–1·01, p=0·0776]), although significantly fewer deaths from stroke (adjusted HR 0·71, 0·53–0·97, p=0·0305) occurred in the amlodipine-based treatment group than in the atenolol-based treatment group. There was no interaction between treatment allocation in the BPLA and in the LLA. However, in the 3975 patients in the non-LLA group, there were fewer cardiovascular deaths (adjusted HR 0·79, 0·67–0·93, p=0·0046) among those assigned to amlodipine-based treatment compared with atenolol-based treatment (p=0·022 for the test for interaction between the two blood pressure treatments and allocation to LLA or not). In the LLA, significantly fewer cardiovascular deaths (HR 0·85, 0·72–0·99, p=0·0395) occurred among patients assigned to statin than among those assigned placebo.
Our findings show the long-term beneficial effects on mortality of antihypertensive treatment with a calcium channel blocker-based treatment regimen and lipid-lowering with a statin: patients on amlodipine-based treatment had fewer stroke deaths and patients on atorvastatin had fewer cardiovascular deaths more than 10 years after trial closure. Overall, the ASCOT Legacy study supports the notion that interventions for blood pressure and cholesterol are associated with long-term benefits on cardiovascular outcomes.
Guidelines for the management of patients with hypertension highlight the importance of blood pressure control, although the target blood pressures, particularly for systolic pressure, remain controversial. Some guidelines advocate preferred drug treatment regimens 1 2 on the basis of the results of cardiovascular disease outcome trials such as the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) 3 and Avoiding Cardiovascular Events through Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) trials, 4 5 whereas others simply focus on blood pressure control irrespective of particular drug classes. 6 7 8 9 10 The trials on which the guidelines are based have typically involved follow-up durations of about 5 years.
Evidence before this study
We searched PubMed from Jan 1, 1990, to May 1, 2018, using the search terms “long term outcomes” or “long term benefits” for clinical trials involving blood pressure-lowering treatment or lipid-lowering with statins. Randomised clinical trials were only included if they reported morbidity or mortality outcomes. For blood pressure, most trials were placebo-controlled and, judging from the single meta-analysis reported so far, the majority were not trials in patients with hypertension (more commonly trials in patients with heart failure and ischaemic heart disease). These trials showed that during a post-trial follow-up of, on average, 41 months, there was 15% reduction in all-cause mortality among patients originally assigned to active drug treatment. To date, the only other trial to have studied long-term outcomes following blood pressure treatment with two or more active treatment groups is ALLHAT. In the case of ALLHAT, the in-trial benefits apparent with the use of chlorthalidone compared with amlodipine and lisinopril, respectively, were no longer evident in the long-term follow-up. A meta-analysis of eight trials of lipid-lowering drugs, including five with statins, reported long-term follow up outcomes, with a significant overall benefit for patients assigned active treatment. Over an average follow-up of 71 months, all-cause mortality was lowered by 10% in patients formerly assigned a statin compared with those assigned placebo. Only two trials reported follow-up longer than 15 years.
Added value of this study
Our findings in patients with hypertension with no previous history of a coronary event show the long-term beneficial effects of antihypertensive treatment with a calcium channel blocker-based treatment regimen and lipid-lowering with a statin. In particular, assignment to amlodipine-based treatment (with perindopril added as required) was associated with a reduction in the number of stroke deaths throughout 16 years of follow-up. Extrapolating our findings from data accrued during the in-trial period, it seems that this benefit is not related to blood pressure differences between treatment regimens. Our findings also add further evidence for the long-term benefits of statin therapy in reducing risk of death from cardiovascular causes. This study is, to our knowledge, the first to report that both blood pressure-lowering and lipid-lowering treatments confer such long-term benefits. Furthermore, our findings from the non-lipid-lowering arm of the trial, a subgroup at higher cardiovascular risk at baseline, showed that these patients also receive significant long-term benefits from blood pressure-lowering therapies.
Implications of all the available evidence
The long-term or legacy benefits of assignment to statins during clinical trials are now well established, with benefits persisting for up to 20 years after the original trials. The mechanisms underlying these observations remain unproven, but it is possible that statin-induced plaque stabilisation occurs during the initial trial, which confers the long-term benefit. The long-term benefit of the calcium channel blocker treatment regimen, which seems to be independent of mean blood pressure levels reported in the original trial, suggests the existence of additional mechanisms for long-term protection. We can only speculate on the nature of these mechanisms, with the most likely being the substantial effects that the calcium channel blocker-based regimen had on blood pressure variability, although other possibilities include the negative effects the atenolol-based regimen had on patients' metabolic profiles. Overall, our findings support the view that interventions on blood pressure and cholesterol are associated with long-term benefits in terms of cardiovascular outcomes.
Previously, hypertension trials comparing active drug treatment with placebo, which have reported substantial post-randomisation (in-trial) blood pressure differences between the two groups, have linked active treatment with long-term benefits. 11 In comparison, long-term post-trial follow-up data from trials comparing active treatments are sparse. 12 It is uncertain, therefore, whether more recent trials that compared active treatment regimens and showed the benefits of a regimen based on a calcium channel blocker and an angiotensin-converting enzyme inhibitor 3 4 would also show a long-lasting beneficial effect. Several long-term follow-up studies of placebo-controlled statin trials have been reported, 13 14 15 which have found persistent benefits in patients previously assigned statin treatment, but none have involved patients with hypertension who were also assigned interventions with different antihypertensive strategies.
ASCOT was designed to compare two antihypertensive treatment strategies and, using a factorial design, to also compare atorvastatin with placebo. 16 In this report, we evaluated the mortality data from the cohort of patients originally recruited into ASCOT from the UK (the ASCOT Legacy study), roughly 16 years after entry into the trial and 10 years after trial closure, to establish whether assignment to either of the original blood pressure- lowering regimens, or to atorvastatin compared with placebo, conferred long-term benefits in terms of all-cause and cause-specific mortality outcomes.
Study design and participants
The detailed ASCOT protocol, including study design, conduct, and baseline characteristics has been published 16 and further detailed information is available on the ASCOT website ). ASCOT was designed to compare two antihypertensive treatment strategies in a blood pressure-lowering arm (BPLA): amlodipine, to which perindopril was added as necessary (amlodipine-based treatment), and atenolol, to which bendroflumethiazide was added as necessary (atenolol-based treatment). Furthermore, using a 2 × 2 factorial design, the trial also compared atorvastatin and placebo in a lipid-lowering arm (LLA) consisting of patients with total cholesterol less than 6·5 mmol/L who were not currently taking a statin or a fibrate.
The study population consisted of men and women with hypertension who were aged 40–79 years at randomisation and had at least three additional risk factors for cardiovascular disease, but no history of coronary heart disease events, currently treated angina, or a cerebrovascular event within 3 months of randomisation. The primary outcome was non-fatal myocardial infarction and fatal coronary heart disease. Patients were originally recruited between Feb 18, 1998, and May 26, 2000, mostly from family practices. In the Nordic countries, individual patients from 686 family practices were enrolled and entered into randomisation, whereas in the UK and Ireland, most patients who underwent randomisation were referred from family practices to 32 regional study centres. In all, 19 257 patients were randomised to the BPLA, of whom 10 240 were randomised to the LLA. There were 9017 patients in the non-LLA group, of whom, about a third were on previous lipid-lowering or aspirin therapy.
In late 2002 and after a median follow-up of 3·3 years, the LLA was stopped prematurely 17 at the recommendation of the data safety monitoring board because of substantial benefits of atorvastatin on the primary endpoint. Patients in the LLA continued to be followed up until the end of the BPLA. During that period, these patients were offered open-label atorvastatin in addition to their assigned blood pressure-lowering treatment. A similar number of patients (approximately two-thirds), formerly assigned either atorvastatin or placebo, went on to receive a statin for the 2·2 years of the LLA extension.
The BPLA was also prematurely stopped at the recommendation of the data safety monitoring board, mainly because of the significantly higher mortality among patients allocated atenolol-based treatment than among those on amlodipine-based treatment. Database lock was in June, 2005, with the last follow-up of patients ranging from December, 2004, to April, 2005. 3 Median follow-up in the BPLA was 5·5 years.
The study conformed to Good Clinical Practice guidelines and the Declaration of Helsinki. The protocol and all subsequent amendments were reviewed and ratified by central and regional ethics review boards in the UK and by national ethics and statutory bodies in Ireland, Sweden, Denmark, Iceland, Norway, and Finland.
ASCOT Legacy cohort
The ASCOT Legacy cohort consisted of all 8580 ASCOT trial participants from the UK. These patients were followed up until the end of the BPLA, during which period 717 died. Of the remaining patients, 7302 (ie, patients from all but two trial sites in the UK from which patient consent for follow-up was not obtained) were flagged with the Office for National Statistics and the General Register Office for Scotland for post-trial follow-up. In this report, we have included all reported deaths up to and including Dec 31, 2015. However, no data on morbidity and treatment were available after the end of the BPLA.
A team of two physicians (JM and AW) independently adjudicated the cause of death, using prespecified criteria that were consistent with the definitions used during the trial period. In these analyses, we report on all-cause mortality, and deaths from cardiovascular and non-cardiovascular causes. All cardiovascular deaths were further adjudicated to report on deaths due to coronary heart disease or stroke. Similarly, non-cardiovascular deaths were sub-categorised to report on cancer-related deaths.
All analyses were done in accordance with the intention-to-treat principle, and thus in-trial follow-up is included for the 561 patients from two UK sites who were not flagged after the closure of the BPLA, and those who dropped out of the BPLA early. For patients who were still alive, censoring was defined as the end of follow-up (Dec 31, 2015) or the end of the BPLA for those who did not consent to long-term follow-up. The end of the LLA period was defined as Oct 1, 2002, and the end of the BPLA was the last follow-up before the database lock in June, 2005.
For both the BPLA and LLA, and for each death outcome, separate Cox proportional hazards models were developed to estimate hazard ratios (HRs) and 95% CIs comparing treatment groups. Both unadjusted and adjusted analyses were done. For the analyses of each cause-specific death, all deaths from other causes were handled as censoring events. We adjusted for the following prespecified covariates at baseline: age, sex, ethnicity, age at leaving full-time education (reflecting socio-economic status), body-mass index (BMI), systolic blood pressure, total cholesterol, presence of diabetes, smoking history, and the other treatment comparison (for example, for the comparisons of the treatment groups in the BPLA, we adjusted for the allocation to statin or a placebo, with a dummy variable for those in the non-LLA group).
For each Cox model, we tested the assumption of proportionality using Schoenfield's residuals, 18 and we found no evidence of any deviation. We prespecified tests for interactions between the two treatment comparisons: ie, between blood pressure lowering treatment regimens and between statin and placebo. Tests for interactions were also done to establish whether the effect of the two blood pressure-lowering treatments differed between subgroups such as those based on the presence of diabetes, age, or allocation to the LLA or not.
Statistical tests were two-sided, with p values less than 0·05 considered statistically significant. All statistical analysis was done using Stata 15.
Role of the funding source
The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
The ASCOT Legacy cohort consisted of 8580 patients from the UK, with a mean age at baseline of 64·1 years (SD 8). Baseline characteristics of these patients were similar to those from Sweden, Denmark, Iceland, Norway, and Finland, 16 except that patients in the UK ASCOT Legacy cohort were more ethnically diverse (10% non-white participants vs 1%), more were men (81·1% vs 72·9%), and fewer were current smokers (23·8% vs 36·1%).
Table 1 shows the baseline characteristics of patients in the ASCOT Legacy cohort. In the BPLA, 4305 patients were assigned to amlodipine-based treatment and 4275 were assigned to atenolol-based treatment. 4605 patients were also included in the LLA, with 2317 randomly assigned to atorvastatin and 2288 to placebo. The remaining 3975 patients comprised the non-LLA group. The baseline characteristics of patients in the LLA and non-LLA groups are shown in the appendix . Compared with the LLA, the non-LLA group contained more women, more patients with more than four cardiovascular risk factors at baseline, and more patients with a history of previous use of lipid-lowering therapy or previous vascular disease, as well as having higher mean baseline total-cholesterol and LDL-cholesterol.
|Blood pressure-lowering arm (n=8580)||Lipid-lowering arm (n=4605)|
|Amlodipine (n=4305)||Atenolol (n=4275)||Atorvastatin (n=2317)||Placebo (n=2288)|
|Age (years)||64 (8)||64 (8)||64 (8)||64 (8)|
|Male sex||3492 (81·1%)||3468 (81·1%)||2016 (87·0%)||2004 (87·6%)|
|White/European||3861 (89·7%)||3840 (89·8%)||2045 (88·3%)||2019 (88·2%)|
|South Asian||130 (3·0%)||109 (2·5%)||72 (3·1%)||80 (3·5%)|
|East Asian||7 (0·2%)||3 (0·1%)||2 (0·1%)||2 (0·1%)|
|Mixed/other||85 (2·0%)||86 (2·0%)||36 (1·6%)||33 (1·4%)|
|African||222 (5·2%)||237 (5·5%)||162 (7·0%)||154 (6·7%)|
|Socioeconomic status (age at leaving full-time education) *|
|12–14 years||1282 (30·0%)||1272 (29·6%)||682 (29·8%)||658 (28·4%)|
|15–16 years||2091 (48·9%)||2165 (50·3%)||1121 (49·0%)||1119 (48·3%)|
|17–18 years||484 (11·3%)||465 (10·8%)||245 (10·7%)||287 (12·4%)|
|>18 years||416 (9·7%)||400 (9·3%)||239 (10·5%)||252 (10·9%)|
|Body-mass index (kg/m 2 )||28·9 (4·7)||28·9 (4·6)||28·8 (4·9)||28·8 (4·6)|
|Current smoker||1035 (24·0%)||1006 (23·5%)||547 (23·6%)||541 (23·6%)|
|Non-drinker||1088 (25·3%)||1089 (25·5%)||574 (24·8%)||571 (25·0%)|
|1–13 units per week||1816 (42·2%)||1831 (42·8%)||1010 (43·6%)||983 (43·0%)|
|≥14 units per week||1401 (32·5%)||1355 (31·7%)||733 (31·6%)||734 (32·1%)|
|Systolic blood pressure (mm Hg)||162 (18)||162 (17)||162 (17)||162 (18)|
|Diastolic blood pressure (mm Hg)||92 (10)||92 (10)||92 (10)||93 (10)|
|Heart rate (beats per minute)||71 (13)||71 (12)||70 (12)||71 (13)|
|Total cholesterol (mmol/L)||5·9 (1·1)||5·9 (1·1)||5·5 (0·8)||5·5 (0·8)|
|HDL-cholesterol (mmol/L)||1·3 (0·4)||1·3 (0·4)||1·3 (0·3)||1·3 (0·3)|
|LDL-cholesterol (mmol/L)||3·8 (1.0)||3·8 (1.0)||3·5 (0.7)||3·5 (0.8)|
|Serum triglycerides (mmol/L)||1·6 (1·2–2·3)||1·6 (1·2–2·3)||1·4 (1·0–2·0)||1·4 (1·1–2·0)|
|Fasting plasma glucose (mmol/L)||5·6 (5·1–6·6)||5·6 (5·1–6·6)||5·6 (5·1–6·5)||5·6 (5·1–6·6)|
|Serum creatinine (μmol/L)||99 (89–109)||98 (89–109)||99 (90–109)||99 (90–109)|
|Presence of diabetes mellitus||1139 (26·5%)||1145 (26·8%)||621 (26·8%)||630 (27·5%)|
|Number of cardiovascular risk factors|
|3 †||2055 (47·7%)||2044 (47·8%)||1201 (51·8%)||1141 (49·9%)|
|4||1416 (32·9%)||1417 (33·1%)||716 (30·9%)||746 (32·6%)|
|≥5||834 (19·4%)||814 (19·0%)||400 (17·3%)||401 (17·5%)|
|History of stroke or transient ischaemic attack (>3 months ago)||507 (11·8%)||492 (11·5%)||233 (10·1%)||239 (10·4%)|
|History of peripheral vascular disease||359 (8·3%)||383 (9·0%)||160 (6·9%)||150 (6·6%)|
|Presence of atrial fibrillation||60 (1·4%)||60 (1·4%)||36 (1·6%)||32 (1·4%)|
|Previous antihypertensive treatment||3961 (92·0%)||3924 (91·8%)||2118 (91·4%)||2106 (92·0%)|
|Previous lipid-lowering treatment||490 (11·4%)||478 (11·2%)||29 (1·3%)||22 (1·0%)|
|Previous aspirin use||1083 (25·2%)||1040 (24·3%)||533 (23·0%)||519 (22·7%)|
* Socioeconomic status was missing for five patients.
† Including 37 who had two risk factors only.Data are n (%), mean (SD), or median (IQR).
Figure 1 shows the numbers of patients and deaths in the ASCOT Legacy cohort by treatment allocation. During a median follow-up of 15·7 years (IQR 9·7–16·4), 3282 (38·3%) of all patients died, including 1640 (38·4%) of 4275 allocated atenolol-based treatment and 1642 (38·1%) of 4305 allocated amlodipine-based treatment. 1768 (53·9%) of all deaths occurred in patients in the LLA, including 903 (39·5%) of 2288 allocated placebo and 865 (37·3%) of 2317 allocated atorvastatin. In total, 1210 (36·9%) deaths were from cardiovascular-related causes.
Table 2 shows the number of events and incidence rates (per 100 person-years) for total and cause-specific mortality with the two blood pressure-lowering treatment regimens for the in-trial period, post-trial period, and throughout all follow-up. During the in-trial period (median 5·5 years follow-up), there were more deaths among patients on atenolol-based treatment than among those on amlodipine-based treatment for all-causes and cardiovascular causes, including the coronary heart disease and stroke components of the cardiovascular mortality. However, during the post-trial period following the closure of the BPLA (an extra 10·7 years of median follow-up), the only treatment differences in mortality were for stroke mortality, for which the differential in the event rates between patients on the two treatment regimens persisted.
|In-trial period (median follow-up 5·5 years)||Post-BPLA period (median follow-up 10·7 years)||Total follow-up (median follow-up 15·7 years)|
|Atenolol-based treatment (n=4275)||Amlodipine-based treatment (n=4305)||Atenolol-based treatment (n=3613)||Amlodipine-based treatment (n=3688)||Atenolol-based treatment (n=4275)||Amlodipine-based treatment (n=4305)|
|Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years|
|Coronary heart disease||86||0·38||66||0·29||127||0·40||132||0·41||213||0·39||198||0·36|
Table 3 and figure 2 show that overall (during a median follow-up of 15·7 years), in the BPLA, there was no significant difference in all-cause mortality between the treatments (HR 0·97, 95% CI 0·90–1·04, p=0·3411). The effect of amlodipine-based versus atenolol-based treatment on cardiovascular death decreased between the in-trial period (0·74, 0·58–0·95, p=0·0177) and total follow-up (0·90, 0·81–1·01, p=0·0776). However, for stroke mortality, the effect of amlodipine-based versus atenolol-based treatment during the in-trial period (HR 0·69, 0·40–1·21, p=0·2013) was similar to that at the end of the follow-up (HR 0·71, 0·53–0·97, p=0·0305), becoming statistically significant with the accrual of more events.
|In-trial period||Total follow-up|
|Unadjusted||Adjusted *||Unadjusted||Adjusted *|
|HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)||p value|
|All-cause mortality||0·93 (0·80–1·07)||0·3130||0·91 (0·78–1·05)||0·2012||0·99 (0·92–1·06)||0·6722||0·97 (0·90–1·04)||0·3411|
|Cardiovascular causes||0·76 (0·60–0·97)||0·0302||0·74 (0·58–0·95)||0·0177||0·93 (0·83–1·04)||0·1909||0·90 (0·81–1·01)||0·0776|
|Coronary heart disease||0·76 (0·55–1·05)||0·0930||0·74 (0·53–1·02)||0·0625||0·92 (0·76–1·11)||0·3786||0·88 (0·73–1·07)||0·2148|
|Stroke||0·69 (0·40–1·21)||0·1969||0·69 (0·40–1·21)||0·2013||0·72 (0·53–0·97)||0·0316||0·71 (0·53–0·97)||0·0305|
|Non-cardiovascular causes||1·04 (0·86–1·25)||0·6965||1·02 (0·85–1·23)||0·8292||1·02 (0·94–1·11)||0·6403||1·01 (0·92–1·10)||0·8880|
|Cancer||1·07 (0·85–1·35)||0·5720||1·06 (0·84–1·34)||0·6101||1·02 (0·91–1·15)||0·7090||1·01 (0·90–1·14)||0·8304|
* Adjusted for baseline age, sex, ethnicity, systolic blood pressure, total cholesterol, body-mass index, diabetes, smoking status, years of education (socioeconomic status), randomisation in the lipid-lowering arm (placebo or atorvastatin) or not.The atenolol-based regimen consisted of atenolol, with thiazide diuretic added as required, and the amlodipine-based regimen consisted of amlodipine, with perindopril added as required. HR=hazard ratio.
There was no interaction between treatment allocation in the BPLA and the LLA ( appendix ). Other subgroup analyses provided no evidence for treatment interactions with either age or diabetes status at baseline (data not shown). However, there were differences in the effect of the two blood pressure treatment regimens, depending on whether a patient was allocated to the LLA or not (p=0·0220 for interaction). Table 4 shows the numbers of events and incidence rates for cause-specific deaths in the two treatment groups, stratified by allocation to the LLA. In the non-LLA group, compared with patients receiving atenolol-based treatment, those allocated to amlodipine-based treatment had significantly fewer cardiovascular deaths (adjusted HR 0·79, 95% CI 0·67–0·93, p=0·0046) and coronary heart disease deaths (adjusted HR 0·76, 0·59–0·93, p=0·0439), and numerically, but not significantly, fewer stroke deaths (adjusted HR 0·67, 0·43–1·04, p=0·0751; Kaplan-Meier plots in the LLA and non-LLA groups are available in the appendix ).
|Atenolol-based treatment||Amlodipine-based treatment||Adjusted hazard ratio (95% CI) *||p value||p interaction †|
|Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years|
|Non-LLA group||769||3·05||745||2·84||0·91 (0·83–1·01)||0·0784||..|
|Non-LLA group||325||1·29||275||1·05||0·79 (0·67–0·93)||0·0046||..|
|Coronary heart disease|
|Non-LLA group||125||0·50||102||0·39||0·76 (0·59–0·99)||0·0439||..|
|Non-LLA group||49||0·19||34||0·13||0·67 (0·43–1·04)||0·0751||..|
|Non-LLA group||444||1·76||470||1·79||1·00 (0·88–1·14)||0·9803||..|
|Non-LLA group||247||0·98||261||1·00||1·01 (0·85–1·20)||0·9436||..|
* Adjusted for age, sex, ethnicity, socioeconomic status, body-mass index, systolic blood pressure, total cholesterol, presence of diabetes, and smoking history.
† p value from test for interaction in adjusted models.The atenolol-based regimen consisted of atenolol, with thiazide diuretic added as required, and the amlodipine-based regimen consisted of amlodipine, with perindopril added as required. The non-LLA group consisted of all patients not included in the LLA because of high baseline total cholesterol. LLA=lipid-lowering arm.
Table 5 shows the numbers of events and incidence rates for cause-specific mortality among patients assigned to atorvastatin or a placebo in the LLA of the ASCOT Legacy cohort, during the in-trial period, post-trial period, and throughout all follow-up.
|In-trial period (median follow-up 3·3 years)||Post-LLA period (median follow-up 13·2 years)||Total follow-up (median follow-up, 15·7 years)|
|Placebo (n=2288)||Atorvastatin (n=2317)||Placebo (n=2198)||Atorvastatin (n=2234)||Placebo (n=2288)||Atorvastatin (n=2317)|
|Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years||Deaths||Incidence per 100 person-years|
|Coronary heart disease||19||0·27||19||0·27||84||0·38||62||0·27||103||0·35||81||0·27|
Table 6 shows estimates for mortality differences between atorvastatin and placebo in the LLA of the ASCOT Legacy cohort during the in-trial period and overall follow-up. Significantly fewer patients assigned to atorvastatin died from cardiovascular causes (HR 0·85, 95% CI 0·72–0·99, p=0·0395) than did those assigned to placebo. Although fewer all cause and coronary heart disease deaths occurred with atorvastatin compared with placebo, differences were not significant (all cause mortality 0·92, 0·84–1·01, p=0·0913 and coronary heart disease mortality 0·78, 0·58–1·04, p=0.0884; figure 3 ). However, there was no evidence of a treatment difference in terms of stroke deaths ( figure 4 ; appendix ). Subgroup analyses by age or presence of diabetes at baseline showed no effect modification (data not shown).
|In-trial period||Total follow-up|
|Unadjusted||Adjusted *||Unadjusted||Adjusted *|
|HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)||p value|
|All-cause mortality||0·92 (0·68–1·24)||0·5970||0·93 (0·69–1·25)||0·6379||0·93 (0·85–1·02)||0·1211||0·92 (0·84–1·01)||0·0913|
|Cardiovascular causes||0·83 (0·51–1·35)||0·4483||0·85 (0·52–1·38)||0·5128||0·85 (0·73–1·00)||0·0459||0·85 (0·72–0·99)||0·0395|
|Coronary heart disease||0·99 (0·53–1·87)||0·9805||1·02 (0·54–1·92)||0·9594||0·77 (0·57–1·03)||0·0735||0·78 (0·58–1·04)||0·0884|
|Stroke||0·75 (0·26–2·17)||0·5995||0·80 (0·27–2·32)||0·6774||1·02 (0·67–1·55)||0·9353||1·02 (0·67–1·55)||0·9238|
|Non-cardiovascular causes||0·99 (0·67–1·44)||0·9393||0·98 (0·67–1·43)||0·9227||0·97 (0·87–1·09)||0·6420||0·96 (0·86–1·08)||0·5440|
|Cancer||1·03 (0·65–1·62)||0·9011||1·01 (0·64–1·59)||0·9598||0·96 (0·82–1·12)||0·5932||0·95 (0·82–1·11)||0·5023|
* Adjusted for age, sex, ethnicity, systolic blood pressure, total cholesterol, body-mass index, diabetes at baseline, smoking status, years of education (socioeconomic status), and randomisation to the blood pressure-lowering treatment.
Our findings from patients with hypertension and no previous coronary events show the long-term benefits of antihypertensive treatment with a calcium channel blocker-based regimen and lipid lowering with a statin. In particular, assignment to amlodipine-based treatment (with perindopril added as required) was associated with fewer stroke deaths throughout 16 years of follow-up ( figure 2 ). Our results also support the long-term benefits of statin therapy in reducing risk of death from cardiovascular causes. 13 14 15 19 20 This study is the first to report that both blood pressure-lowering and lipid-lowering treatments confer such long-term benefits. Furthermore, findings for the higher risk subgroup, the non-LLA group, supported the long-term benefits of blood pressure lowering therapies in such patients. Additionally, the benefits conferred differed between the blood pressure lowering regimens, despite the in-trial gain in blood pressure control being similar ( appendix )
The only other large trial to have studied cardiovascular outcomes with both antihypertensive therapy and lipid- lowering with statins, the Antihypertensive and Lipid Lowering to prevent Heart Attacks Trial (ALLHAT), 21 22 compared different monotherapies with similar add-on antihypertensive treatments. Both ASCOT and ALLHAT assessed the potential benefits of statins in patients with hypertension. In ASCOT, atorvastatin was compared with placebo, whereas in ALLHAT, pravastatin was compared with usual care. Unfortunately, in ALLHAT, many patients in the usual care group received statins and only a small difference in cholesterol was detected between the treatment groups, which resulted in the trial being underpowered to compare effects on major cardiovascular endpoints. 22 23 The findings from the extended follow-up of patients assigned to three arms of the antihypertensive study, the chlorthalidone, amlodipine, and lisinopril arms, were mixed. 24 25 In-trial benefits that were apparent with the use of chlorthalidone versus amlodipine and lisinopril, respectively, were no longer evident in the long-term follow-up.
In long-term follow-up of trials in patients with hypertension 11 that compared active treatment with placebo, and in which blood pressure differences were associated with substantial reductions in cardiovascular events, a legacy or carryover effect has been seen in the post-trial period, 26 27 28 29 with long-term reductions—on average 9%—in mortality in the groups previously receiving active treatment. 11 However, in ASCOT, which compared different active (alternative) treatment strategies, 16 the long-term benefits associated with the amlodipine-based regimen could not be attributed to earlier differences in blood pressures during the trial. 30 First, as previously reported with the main trial outcome 3 and in a post-hoc analysis, 30 there was only a small difference in blood pressure of 2·9/1·8 mm Hg between the two treatment groups. Furthermore, in the ASCOT UK Legacy population, overall mean differences between blood pressure-lowering regimens in blood pressure recorded during the trial were only around 1·2/1·6 mm Hg ( appendix ). This small difference in blood pressure cannot account for the sustained and significant differences in stroke mortality apparent in this long-term follow-up.
Initially, we attributed the benefits of the amlodipine-based treatment to differences in the metabolic profile, including adverse effects of glycaemia associated with the atenolol-based therapy, and other differences including small changes in lipids and electrolytes. 30 However, these changes alone were insufficient to explain the observed in-trial effects on mortality and cardiovascular events. 30 We have subsequently described other important differences between the two treatment arms, with atenolol-based treatment lowering central aortic pressure substantially less than amlodipine-based treatment. 31 Additionally, blood pressure variability was a major determinant of cardiovascular outcome in ASCOT, 32 and was reduced to a much greater extent with the amlodipine-based treatment compared with the atenolol-based regimen. 33
We believe that these mechanisms, and potentially others that are currently unknown, are possible explanations for the findings in ASCOT-BPLA, and that they also contribute to the long-term benefits we identified in the ASCOT Legacy population. We are in the process of investigating these mechanisms further, together with studies on potential genetic and other biomarker predictors of cardiovascular events, which might explain the differences we saw during this 16 year follow-up.
For the blood pressure trial, the most striking treatment difference between regimens was in stroke death, which, according to our earlier reports, 33 was closely associated with blood pressure variability. Our observations also show that, although there was a significant reduction in cardiovascular mortality with amlodipine versus atenolol during the trial, this effect was attenuated in subsequent follow-up, except for stroke mortality, where the beneficial effect remained essentially the same even 10 years after trial closure. These dilutional effects might reflect that, after the trial, both groups of patients would have received similar overall treatment strategies. Nevertheless, the persistence of the effect size for stroke deaths might reflect the close relationship that this outcome has with blood pressure treatment. 34
We also observed a differential legacy effect between the two treatment groups, according to the baseline risk of the patients. In our evaluation of event rates between patients included in the LLA and those in the non-LLA group ( table 4 ), among those assigned amlodipine-based treatment, there was little difference in stroke mortality and cardiovascular mortality and very small differences in coronary heart disease mortality. These findings suggest that amlodipine-based treatment confers a similar effect irrespective of baseline risk. However, for patients assigned atenolol-based treatment there was, compared with those in the LLA, an increased event rate among those in the non-LLA group.
Following the early stopping of the LLA, the beneficial effects of atorvastatin on mortality persisted, with little dilution of the effect during the long-term follow-up. In the ASCOT-LLA Legacy cohort, the effect sizes of cardiovascular and all-cause mortality at 13 years remained similar to those at the end of the trial, but, with an increasing number of events, the mortality benefits became significant for cardiovascular deaths, with non-significant suggestions of long-term benefit for all cause and coronary heart disease mortality. Similar findings, particularly the relationship between the use of a statin and reduction in cardiovascular mortality in long-term follow-up have been reported in other trials. 13 14 15 19 20 The mechanism for these durable and legacy effects remains unclear, but it is possible that statin-induced plaque stabilisation occurs during the initial trial period, which confers a long term benefit on cardiovascular outcomes.
The present analyses have several limitations. After the closure of the trial, we have no data on antihypertensive and lipid-lowering medications, or indeed on any other treatments. Thus, we cannot reliably ascertain what differences, if any, in post-trial blood pressures and treatments existed in the longer term. However, after the closure of the LLA, a similar number of patients previously assigned to atorvastatin or placebo received a statin during the 2·2 year extension during final years of the BPLA of the trial. 35 Another limitation is that we have no morbidity data after trial closure. The availability of such data would allow for a more comprehensive evaluation, and we are in the process of developing appropriate registry linkages to acquire them. Furthermore, the generalisability of the study population could be questioned. The patients in this cohort had a median age of 64 years, they were hypertensive, and they had several common risk factors, which would make them representative of individuals of a similar age in the community. Indeed, in previous unpublished work, we have compared our population with the patients with hypertension from the community, as reported in the Health Survey of England, and we found them to be similar. Lastly, our findings need to be replicated in other studies. The greatest strength of this study is that it is the first to report on a large cohort of patients with hypertension and investigate both blood pressure-lowering and lipid-lowering treatments and their effects on long-term mortality, with substantial power to evaluate differences in mortality between treatments.
The legacy outcomes from the ASCOT trial show the long-term benefits of antihypertensive treatment with a calcium channel blocker-based treatment regimen and lipid-lowering with a statin, 16 years after entry into the trial and more than 10 years after its closure. To our knowledge, we report for the first time that the legacy benefit from the amlodipine-based regimen in terms of reducing risk of stroke mortality seems to be independent of blood pressure levels achieved. Our data contribute to evidence to support the long-term benefits of statins for reducing cardiovascular mortality. Overall, our findings support the notion that interventions for blood pressure and cholesterol are associated with long-term benefits for cardiovascular outcomes.
This online publication has been corrected. The corrected version first appeared at thelancet.com on October 18, 2018
AG and PS had the original idea for the Legacy study, designed the protocol, supervised the data collection and statistical analyses, and wrote the initial drafts of the manuscript. JM and AW did the independent and blinded classification of all deaths and commented on the manuscript. Statistical analyses were done by AG, TG, and TC. SP provided statistical oversight and reviewed the manuscript. NP contributed to protocol design and reviewed the manuscript.
Declaration of interests
AG reports support from the Foundation for Circulatory Health, during the conduct of the study, and non-financial support from Servier, outside the submitted work. NP reports grants from Pfizer and Servier, during the conduct of the study; grants and other support from Servier for serving on advisory boards, speaking at educational meetings, and conducting the PREMIUM trial; and other support from AstraZeneca and Napi for speaking at educational meetings, outside the submitted work. Additionally, NP is President of the International Society of Hypertension. PS reports grants and Personal fees from Pfizer and Servier, during the conduct of the study, and grants and personal fees from Pfizer, Servier, and Amgen, outside the submitted work. All other authors declare no competing interests.
- 1 T Krause, K Lovibond, M Caulfield, T McCormack, B Williams. Management of hypertension: summary of NICE guidance. BMJ. 2011;343:d4891
- 2 RJ McManus, M Caulfield, B Williams. NICE hypertension guideline 2011: evidence based evolution. BMJ. 2012;344:e181
- 3 B Dahlof, PS Sever, NR Poulter, et al. Prevention of cardiovascular events with an antihypertensive regimen of amlodipine adding perindopril as required versus atenolol adding bendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomised controlled trial. Lancet. 2005;366:895-906
- 4 K Jamerson, MA Weber, GL Bakris, et al. Benazepril plus amlodipine or hydrochlorothiazide for hypertension in high-risk patients. N Engl J Med. 2008;359:2417-2428
- 5 GL Bakris, PA Sarafidis, MR Weir, et al. Renal outcomes with different fixed-dose combination therapies in patients with hypertension at high risk for cardiovascular events (ACCOMPLISH): a prespecified secondary analysis of a randomised controlled trial. Lancet. 2010;375:1173-1181
- 6 PA James, S Oparil, BL Carter, et al. 2014 evidence-based guideline for the management of high blood pressure in adults: report from the panel members appointed to the Eighth Joint National Committee (JNC 8). JAMA. 2014;311:507-520
- 7 G Mancia, R Fagard, K Narkiewicz, et al. 2013 ESH/ESC guidelines for the management of arterial hypertension. Eur Heart J. 2013;34:2159-2219
- 8 AA Leung, K Nerenberg, SS Daskalopoulou, et al. Hypertension Canada's 2016 Canadian hypertension education program guidelines for blood pressure measurement, diagnosis, assessment of risk, prevention, and treatment of hypertension. Can J Cardiol. 2016;32:569-588
- 9 D Stephan, S Gaertner, EM Cordeanu. A critical appraisal of the guidelines from France, the UK, Europe and the USA for the management of hypertension in adults. Arch Cardiovasc Dis. 2015;108:453-459
- 10 PK Whelton, RM Carey, WS Aronow, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Hypertension. 2018;71:e13-e115
- 11 WJ Kostis, L Thijs, T Richart, JB Kostis, JA Staessen. Persistence of mortality reduction after the end of randomized therapy in clinical trials of blood pressure-lowering medications. Hypertension. 2010;56:1060-1068
- 12 EK Chowdhury, A Owen, Z Ademi, et al. Short- and long-term survival in treated elderly hypertensive patients with or without diabetes: findings from the Second Australian National Blood Pressure study. Am J Hypertens. 2014;27:199-206
- 13 WJ Kostis, AE Moreyra, JQ Cheng, JM Dobrzynski, JB Kostis. Continuation of mortality reduction after the end of randomized therapy in clinical trials of lipid-lowering therapy. J Clin Lipidol. 2011;5:97-104
- 14 I Ford, H Murray, C McCowan, CJ Packard. Long term safety and efficacy of lowering LDL-cholesterol with statin therapy: 20-year follow up of West of Scotland Coronary Prevention Study. Circulation. 2016;133:1073-1080
- 15 WE Hague, J Simes, A Kirby, et al. Long-term effectiveness and safety of pravastatin in patients with coronary heart disease: 16 years of follow-up of the LIPID Study. Circulation. 2016;133:1851-1860
- 16 PS Sever, B Dahlöf, NR Poulter, et al. Rationale, design, methods and baseline demography of participants of the Anglo-Scandinavian Cardiac Outcomes Trial. ASCOT investigators. J Hypertens. 2001;19:1139-1147
- 17 PS Sever, B Dahlöf, NR Poulter, et al. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower-than-average cholesterol concentrations, in the Anglo-Scandinavian Cardiac Outcomes Trial—Lipid Lowering Arm (ASCOT-LLA): a multicentre randomised controlled trial. Lancet. 2003;361:1149-1158
- 18 D Schoenfeld. Partial residuals for the proportional hazards regression-model. Biometrika. 1982;69:239-241
- 19 TE Strandberg, K Pyorala, TJ Cook, et al. Mortality and incidence of cancer during 10-year follow-up of the Scandinavian Simvastatin Survival Study (4S). Lancet. 2004;364:771-777
- 20 Heart Protection Study Collaborative Group. Effects on 11-year mortality and morbidity of lowering LDL cholesterol with simvastatin for about 5 years in 20 536 high-risk individuals: a randomised controlled trial. Lancet. 2011;378:2013-2020
- 21 The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic: the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT). JAMA. 2002;288:2981-2997
- 22 The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major outcomes in moderately hypercholesterolemic, hypertensive patients randomized to pravastatin vs usual care: the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT-LLT). JAMA. 2002;288:2998-3007
- 23 KL Margolis, BR Davis, C Baimbridge, et al. Long-term follow-up of moderately hypercholesterolemic hypertensive patients following randomization to pravastatin vs usual care: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT-LLT). J Clin Hypertens. 2013;15:542-554
- 24 WC Cushman, BR Davis, SL Pressel, et al. Mortality and morbidity during and after the antihypertensive and lipid-lowering treatment to prevent heart attack trial. J Clin Hypertens. 2012;14:20-31
- 25 JM Yamal, S Oparil, BR Davis, et al. Stroke outcomes among participants randomized to chlorthalidone, amlodipine or lisinopril in ALLHAT. J Am Soc Hypertens. 2014;8:808-819
- 26 N Shulman, E Tuttle, G Entwisle, et al. Persistence of reduction in blood pressure and mortality of participants in the hypertension detection and follow-up program. JAMA. 1988;259:2113-2122
- 27 JB Kostis, AC Wilson, RS Freudenberger, NM Cosgrove, SL Pressel, BR Davis. Long-term effect of diuretic-based therapy on fatal outcomes in subjects with isolated systolic hypertension with and without diabetes. Am J Cardiol. 2005;95:29-35
- 28 JA Staessen, L Thijisq, R Fagard, et al. Effects of immediate versus delayed antihypertensive therapy on outcome in the Systolic Hypertension in Europe Trial. J Hypertens. 2004;22:847-857
- 29 MR Nelson, EK Chowdhury, J Doust, CM Reid, LM Wing. Ten-year legacy effects of baseline blood pressure ‘treatment naivety’ in the Second Australian National Blood Pressure study. J Hypertens. 2015;33:2331-2337
- 30 NR Poulter, H Wedel, B Dahlöf, et al. Role of blood pressure and other variables in the differential cardiovascular event rates noted in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA). Lancet. 2005;366:907-913
- 31 B Williams, PS Lacy, SM Thom, et al. Differential impact of blood pressure-lowering drugs on central aortic pressure and clinical outcomes: principal results of the Conduit Artery Function Evaluation (CAFE) study. Circulation. 2006;113:1213-1225
- 32 PM Rothwell, SC Howard, E Dolan, et al. Prognostic significance of visit-to-visit variability, maximum systolic blood pressure, and episodic hypertension. Lancet. 2010;375:895-905
- 33 PM Rothwell, SC Howard, E Dolan, et al. Effects of beta blockers and calcium-channel blockers on within-individual variability in blood pressure and risk of stroke. Lancet Neurol. 2010;9:469-480
- 34 D Ettehad, CA Emdin, A Kiran, et al. Blood pressure lowering for prevention of cardiovascular disease and death: a systematic review and meta-analysis. Lancet. 2016;387:957-967
- 35 PS Sever, NR Poulter, B Dahlöf, et al. The Anglo-Scandinavian Cardiac Outcomes Trial lipid lowering arm: extended observations 2 years after trial closure. Eur Heart J. 2008;29:499-508
© 2018 Elsevier Ltd.
© 2018 Elsevier Ltd. All rights reserved.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds or experiments described herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. To the fullest extent of the law, no responsibility is assumed by Elsevier for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.
This e-print is distributed with the support of Pfizer.
Elsevier España, S.L.U.
(A member of Elsevier)
Av. Josep Tarradellas, 20-30
Tel.: 932 000 711
Fax: 932 091 136